Skip to content
SDGs Center UDINUS

Evidence record · UDINUS · SDGs

Hyperparameter Tuning with Optuna to Optimize the YOLOv11n Model for Weed Detection

Rektorat UDINUS 2025 Published 2026-10-30

Evidence record · UDINUS · SDGs

Unit
Rektorat UDINUS
Period
2025
Goals
Record
1da7396c-ccd3-4b3d-8fa8-83e97c80430c
Published
2026-10-30

Universitas Dian Nuswantoro (UDINUS), through its researchers Arsyad, Candhy Fadhila; Andono, Pulung Nurtantio; and Soeleman, Moch Arief, conducted a study on hyperparameter tuning with Optuna to optimize the YOLOv11n model for weed detection. The activity addressed the importance of accurate weed detection for maintaining the cleanliness and aesthetic appeal of residential yards. It also examined the use of a lightweight object detection model in real-world conditions, with attention to computational efficiency for resource-limited devices such as drones and IoT systems.

The study applied Optuna as an automatic hyperparameter optimization framework. The research used data augmentation techniques comprising crop with 0–20% zoom, hue adjustment of ±20°, saturation adjustment of ±30%, brightness adjustment of ±20%, exposure adjustment of ±15%, and mosaic augmentation. The augmented images were used to train four YOLO nano variants, namely YOLOv5n, YOLOv8n, YOLOv11n, and YOLOv12n. The models were evaluated using Precision, Recall, F1-Score, and mean Average Precision (mAP), providing a basis for comparing their weed-identification performance.

Among the models tested, YOLOv11n with the Custom Optuna configuration delivered the highest performance, achieving a 94.6% F1-score and 97.8% mAP@0.5. The findings demonstrate that the optimized YOLOv11n model can support accurate and efficient real-time weed detection in household environments, particularly on edge devices with limited hardware capabilities. The study therefore presents an activity related to the second goal through research addressing a practical agricultural and gardening application, while remaining focused on model optimization, weed identification, and computational efficiency. UDINUS frames this work as research conducted by its researchers to develop and evaluate an optimized detection approach for practical implementation in precision agriculture and smart gardening.

Reference https://www.scopus.com/pages/publications/105022252020?origin=resultslist